Do we see apples as edible? (стр. 1 )

Do we see apples as edible?

Bence Nanay

Do we (sometimes) perceive apples as edible? One could argue that it is just a manner of speaking to say so: we do not really see an object as edible, we see it as having certain shape, size and color and we only infer on the basis of these properties that it is. I argue that we do indeed see objects as edible, and do not just believe that they are. My argument proceeds in two steps. First, I point out that Susanna Siegel’s influential argument in favor of the claim that we represent sortal properties perceptually does not work. Second, I argue that we can fix this argument if we replace the sortal property in question with the property of being edible, climbable or Q-able in general.

I. Introduction

We perceive objects as having various properties. We perceive them as having a certain color, a certain shape and a certain spatial location, for example. The question I would like to examine here is whether we perceive them as having a kind of property that is less obviously perceptual: the property of being edible, climbable or Q-able in general.

The first thing to note is that there are two ways of asking this question that correspond to two different ways of thinking about perceptual experiences. According to some, perceptual experiences represent objects as having various properties.[1] If we go along with this assumption, the question will be: what properties are represented by perceptual experiences. But according to others, perceptual experiences do not represent anything: they are not representations at all.[2] They may be thought of as ‘presenting’ or ‘being sensitive’ to some properties, but they do not represent these properties. If we accept this framework, the question is what properties perceptual experiences are sensitive to. The important point is that regardless of whether we accept a ‘representational’ or a ‘relational’ view of perceptual experience,[3] the question about which properties are perceptually experienced and which ones aren’t can (and should) be raised. In what follows, I will use the ‘representational’ terminology and assume that perceptual states represent objects as having various properties, but the argument can be rephrased to fit the relationalist framework.

Most of the discussion about what properties are represented by perceptual experiences concerns sortal properties. When I am looking at the object in front of me do I perceive it as a table or do I perceive it as having a certain shape, size and color and I only infer that it is a table? In other words, besides the properties of having certain shape, size and color, is the property of being a table also represented in perception?

I will not take sides in this debate, as my aim is to show that properties that could be thought of as even less obviously perceptual are represented in perception: the property of being edible, climbable or Q-able in general.

But I will use some important considerations and arguments from the debate about the perceptual representation of the property of being a table. The structure of my argument is the following. First, I try to show that Susanna Siegel’s influential argument in favor of the claim that we represent the property of being a table perceptually does not work. Second, I argue that we can fix this argument if we replace the sortal property in question with the property of being Q-able.

The claim that we perceive objects as Q-able is not new. I myself argued recently that our perceptual system represents objects as edible, climbable or Q-able in general.[4] My main concern there, however, is to establish that our perceptual system represents properties of this kind – consciously or unconsciously. For the purposes of that argument, explicitly remained neutral about whether these representations, which I label ‘action-oriented perceptual states’ are conscious.[5] The present paper, in contrast, is about perceptual experiences. The question is whether we are perceptually aware of properties like edible, climbable or Q-able in general. Everything I say in this paper is about perceptual experiences: when I talk about ‘seeing’ in what follows, that is to be understood as ‘consciously seeing’ (much like the concept is used by Fred Dretske.[6] The argument in the present paper supports my earlier, more general claim, but it goes further: it aims to make a stronger claim: at least sometimes, we consciously perceive objects as edible.[7]

II. Are sortal properties represented in perception?

Susanna Siegel argues that some sortal properties (she calls them K-properties), like being a pine tree or being a table, are represented in perception. In this section, I will try to show that her argument is vulnerable to an important objection. Then I will try to fix Siegel’s argument in such a way that it does show that some not obviously perceptual properties, like that of being Q-able, are represented in perception.

Siegel’s argument is the following. Take two experiences, E1 and E2: the experience of a tree before and after taking a course on the typology of trees, respectively. Here is what we can say about these two experiences:

(0) The overall experience of which E1 is a part differs from the overall phenomenology of which E2 is a part.

(1) If the overall experience of which E1 is a part differs from the overall phenomenology of which E2 is a part, then there is a phenomenal difference between the sensory experiences E1 and E2.

(2) If there is a phenomenal difference between the sensory experiences E1 and E2, then E1 and E2 differ in content. 

(3)  If there is a difference in content between E1 and E2, it is a difference with respect to sortal properties represented in E1 and E2.

-------------------------------------------------------------------------------------------------

(Conclusion) Sortal properties are represented in perception.

So the overall experience of seeing a pine tree before learning about pine trees is different from the overall experience after having learned this (0). This difference is perceptual difference (1) and it is due to the difference of the properties represented in perception in the two cases (2). Finally, this difference with regards to the represented properties is a difference with regards to the represented sortal properties.

I accept premise (0), (1) and (2), for the sake of argument – I will have more to say about them in Section VI. I will argue against (3).

Denying (3) (while accepting (2)) would amount to saying that E2 represents some (non-sortal) property, whereas E1 does not and this explains the difference in phenomenology without any need to talk about sortal properties.

Siegel considers a version of this way of arguing against (3). The suggestion is that E1 and E2 differ with respect to the Gestalt-properties (i. e., complex shape, size and color-properties) they represent. E1 represents some Gestalt-property and E2 represents some other Gestalt-property. Neither of them represents any sortal properties. Siegel admits that this explanatory scheme may work in some cases, but she argues that it does not work in all cases.

She gives the following example for a case where the appeal to Gestalt-properties breaks down. X has an unusual facial expression when he is expressing doubt. When I first saw him making this face, I had no idea that he was expressing doubt. But as I got to know him, I learned to recognize his doubtful expression. E1 is my experience of X’s face before I knew it was expressing doubt and E2 is my experience after I came to know his expression. Siegel assumes, rightly, that there may be a phenomenal difference between E1 and E2. Importantly, she claims that:

it seems implausible to suppose that there must be a change in which color and shape properties are represented before and after one learns that it is doubt that the fact so contorted expresses.[8]

The structure of Siegel’s argument is then the following. There are cases where all the following are true:

(a) E1 and E2 represent the same Gestalt-properties.

(b) Still, there is a phenomenal difference between E1 and E2.

(c) The phenomenal difference between E1 and E2 must be a representational difference (this follows from step (2) above).

----------------------------------------------------------------------------

(Conclusion) Hence, E1 and E2 must represent different sortal properties

I will not question (b): it does seem that there is a phenomenal difference between seeing X’s doubtful expression before and after learning that it expresses doubt. I have already accepted (c). But I will try to point out that (a) is ambiguous and that it does not seem to be a plausible assumption if we do some disambiguating.

III. The importance of attention

We attend to some, but not all, the properties we represent objects as having. Some properties we represent without attending to them, but some others we represent and attend to.[9] As William James famously wrote, "attention […] out of all the sensations yielded, picks out certain ones as worthy of notice and suppresses all the rest. We notice only those sensations which are signs to us of things which happen practically […] to interest us".[10] To put it very simply, some properties are represented pre-attentively and a subset of these properties is also represented post-attentively.[11]

In the light of this distinction, premise (a) of Siegel’s argument can mean one of the following two claims:

(a1) E1 and E2 pre-attentively represent the same Gestalt-properties.

(a2) E1 and E2 post-attentively represent the same Gestalt-properties.

My worry is that while we have good reason to believe that (a1) is true, what would be needed for Siegel’s argument to work is (a2). But it is far from clear that we have any reason to hold (a2).

I will not question claim (a1), which seems convincing enough. In the case of E1 and E2, we have the same object in front of us, and this object has the very same properties. Our perceptual apparatus is also sensitive to the same properties in the two cases. Hence, it seems reasonable to say that E1 and E2 represent the same Gestalt-properties pre-attentively. But it would be a much stronger claim to say that they represent the same Gestalt-properties post-attentively.